Why Branded Search Is the Least Understood Line Item on a Med-Spa's Ad Bill

The line that reports the best return is the one to read first

Open almost any local-service Google Ads account and the branded campaign, the one bidding on the practice's own name, looks like the star. Its click-through rate is high, its cost per click is low, its conversion rate is the envy of every other campaign, and its reported return on ad spend is often several times higher than non-branded search. On a med-spa's monthly bill it reads as the safest money on the page.

That appearance is exactly why it is the least understood line item. Every other campaign has to earn a click from a stranger. The branded campaign intercepts someone who already typed your name, which means they already knew you, already intended to find you, and in most cases would have reached you through the free organic listing that sits directly below the ad. The metrics do not distinguish the customer the ad created from the customer it merely collected. Read as reported, branded search takes credit for demand it did not generate.

The eBay experiment: what branded search actually caused

The load-bearing evidence here is not a vendor case study. It is a peer-reviewed field experiment. In work published in Econometrica, economists Thomas Blake, Chris Nosko, and Steven Tadelis ran a large-scale randomized test at eBay, switching paid search on and off across markets to measure what the ads truly caused rather than what a dashboard reported.

For branded keywords, searches for eBay's own name, the result was stark: turning the ads off produced no measurable short-term drop in sales. Nearly all the traffic the branded ads had been paying for simply arrived through the organic ("natural") search result instead. The spend was buying clicks that were already going to happen.

The non-branded finding was subtler and just as important. New and infrequent users could be positively influenced by ads, but frequent users, the people who were going to buy regardless, absorbed most of the spend, and their purchases were unaffected. Averaged across everyone, returns for the tested non-brand campaigns ran negative. The single mechanism under both results is consumer heterogeneity: the same ad has a very different effect on someone discovering you than on someone already walking through your door.

Why the null result travels to a med-spa, and where it does not

A word of caution before the analogy does any work: the eBay study is one firm, a giant marketplace, not a local aesthetics practice, and its authors are careful to frame it as a documented mechanism rather than a universal constant. We apply it as a lens, not as a measured figure for your account, and any specific dollar of med-spa waste has to be established with your own data, not borrowed from eBay's.

What travels is the mechanism, not the number. A branded search on "[Your Practice Name] med spa" is the most intent-saturated query in your whole account. The person typing it is not comparison shopping; they are navigating to you. That is precisely the condition under which the eBay experiment found paid clicks to be substitutes for free organic clicks rather than additions to them. If your practice already owns the top organic result for its own name, and for a distinctive brand almost every practice does, the branded ad is frequently paying to sit above a free listing the same searcher would have clicked.

What does not travel automatically is the conclusion "so turn it off." The eBay account had no competitor actively hijacking its brand term in the way a local market can. That exception is where branded search stops being a tax and becomes a shield, which is the next section.

Should I bid on my own brand name? When it is protection, not waste

It depends on who else is in the auction. Branded search earns its place when a specific, checkable condition holds, and wastes money when it does not.

Sponsored search sits inside a generalized second-price auction, the mechanism Edelman, Ostrovsky, and Schwarz documented as the engine behind essentially all keyword advertising. Nothing stops a rival from entering that auction on your name. When a competing med-spa or a national chain bids on your brand term, the top of the page for your own name can be handed to them unless you are present to hold it. In that case the branded ad is not buying incremental discovery; it is denying a competitor a cheap interception of buyers who were looking for you specifically. That is defense, and it can be worth paying for.

The measurement trap: reported return versus caused return

Branded search flatters itself for a reason rooted in how the numbers are formed. The platform attributes a booking to the last ad clicked, and the branded ad is very often the last click before a customer who was always going to convert. This is a specific case of a broader, well-documented problem.

Randall Lewis, Justin Rao, and David Reiley called it activity bias: across three controlled experiments, they showed that people who are already active online, browsing, searching, clicking, are more likely to do all of those things at once, ad or no ad. Correlation between ad exposure and conversion gets read as causation, and effectiveness is systematically overestimated. Branded search is where this bias bites hardest, because ad exposure and buying intent are almost perfectly correlated there by construction.

Industry practitioners put a range on the gap. Measured incremental return on ad spend is commonly reported to run well below platform-reported return, with branded search named as the channel where the divergence is worst and estimates that a large majority of branded clicks would have converted through organic or direct traffic anyway. We flag those specific percentages as industry-reported and unaudited, directionally consistent with the peer-reviewed experiments above but not a substitute for them, and not a number we would attach to your account without measuring it ourselves.

How to actually find out, for your own account

The reason branded search stays misunderstood is that the dashboard cannot answer the only question that matters: what would happen to bookings if this line turned off. Answering it requires a test that creates a counterfactual, and there are three ways to build one, in rising order of rigor.

A field guide to reading your med-spa's branded line

Pulling the evidence into practice, here is how to read the branded line item the next time the invoice lands. None of this requires you to accept a number on faith; each step replaces an assumption with a check.

• Treat the reported return on branded search as an upper bound, never as the caused return. The true figure is lower by an amount only a test can reveal.
• Separate the branded campaign in your reporting so its spend and conversions never blend into the non-branded numbers. You cannot audit a line you cannot see.
• Run the competitor check before touching the budget. Live competitor bids on your name justify defense; an empty auction does not.
• Watch total booked appointments, the real outcome, not platform conversions, when you test. Paid conversions can fall while bookings hold steady, which is the whole point.
• Med-spa keywords are expensive. Category terms in this vertical are among the priciest in local service, so a branded campaign quietly wandering into non-branded queries wastes money quickly. The specific cost-per-click figures for your market need your own data, not an industry average.

The point is not to cut, it is to know

The failure mode this article argues against is not "spending on your brand name." It is spending on it without knowing which kind of spend it is. Branded search can be a genuine shield against competitors interposing themselves at the moment a buyer is looking for you by name. It can equally be a silent tax on demand you already own, dressed up by last-click attribution as your best-performing campaign.

The difference is not a matter of opinion or of a vendor's confidence. It is measurable, and the tools to measure it are decades old and peer-reviewed. The businesses that stop overpaying for their own name are the ones that treated the branded line as a question to be tested rather than a result to be trusted.